Variable names

Allowed names for identifier (name of variable or function):

• Any sequence of alphabetical character, number, or underline character _

• First symbol has to be a character
• Case sensitive
• Not a python built-in keyword

Name space

The Python interpreter maps the name of an identifier in the active name space. The active name space depends on the active code block (module, function). The identifier can be local or global with respect to the active code block.

   1 def bad_style(x):
   2         # Name space of bad_style
   3         global y
   4         y=x 
   5         z=x
   6         return x
   7 
   8 # Name space of __main__
   9 x,y,z=2,2,2
  10 x=bad_style(1)
  11 print 'x=',x,'y=',y,'z=',z
  12 # y is changed!

   1 import scipy
   2 scipy.pi # pi is in the name space of scipy
   3 
   4 from scipy import pi
   5 pi # pi is in the active name space
   6 
   7 from scipy import *
   8 pi # all identifiers of scipy are in the active name space
   9 
  10 del(pi) # Deletes pi from the active name space
  11 reset   # ipython function to delete all identifiers

Built-in datatypes

Scalar

• Plain integer: 9

• Long integer: 9**99, 1L

• Hex integer: 0x10

• Floating point: 0.1

• Exponential floating point: 1e-3

• Complex: 3+2j

Casting

• int()
• long()
• float()
• complex()
• str()

Sequences

String

   1 s1='Hello world'
   2 s2="Hello world"
   3 s1==s2
   4 
   5 s='"'
   6 s="'"

Dictionary

   1 D={1:'one',2:'two','one':1,'two':2}
   2 D['one']
   3 D[1]

Extended data types

Array

NumPy/SciPy provides a multidimensional array data type. An array can hold arbitrary Python objects but usually they are used for N-dimensional numeric data types.

Array creation

• empty((d1,d2),dtype) returns uninitialized array of shape d1,d2.

• zeros((d1,d2),dtype) returns array of shape d1,d2 filled with zeros

• ones((d1,d2),dtype) returns array of shape d1,d2 filled with ones

• array(object,dtype) returns an array from an object, e.g. a list

• dtype fundamental C data type e.g. uint8, int16, int64, float32, float64

Array indexing

• A[y,x] returns (y,x) element of the array

• A[:,x] returns all elements of the y-dimension at x

• A[y1:y2,x]

• A[:,:] returns a copy of two dimensional array A

• A[:,:,0] returns the first sub-image of a 3-dimensional array

Functions

Function with one argument

In [124]: def quadrat(x):
   .....:     return x**2
   .....:

In [125]: quadrat(2)
Out[125]: 4

Function with variable number of arguments

In [126]: def printargs(*args):
   .....:     for arg in args:
   .....:         print arg
   .....:

In [127]: printargs(a)
[1, 2, 3]

In [128]: printargs(a,range(4))
[1, 2, 3]
[0, 1, 2, 3]

Function with variable number of keywords

In [131]: def printkeys(**kw):
   .....:     for k in kw.keys():
   .....:         print k,kw[k]
   .....:

In [132]: printkeys(key1=10,key2='Hallo')
key2 Hallo
key1 10

Function with functions as arguments

Functions can be passed to functions as arguments.

   1 from scipy import sin,pi
   2 
   3 def func(x,f):
   4         return f(x)
   5 
   6 print func(pi/2,sin)

eval(), apply()

In Matlab or IDL the passing of functions to functions works only with a detour using feval() or Call_function()

• IDL passing functions to functions

Of course, this way is possible in Python as well using the eval() or apply() function.

   1 apply(sin,(pi/2,))
   2 
   3 f='sin'
   4 x=pi/2.0
   5 eval(f+'('+str(x)+')')

In Fortran it is also possible to pass functions as arguments but with quite some overhead:

• http://wingolog.org/software/fortran/tips-tricks.html

Lambda expressions

Lambda expressions are (anonymous) one line functions

   1 l=lambda x: x**2-1
   2 l(4)

Mapping

Similar as apply() works with one argument one can call map() with a sequence

   1 map(len,['Always','look','on','the','bright','side','of','life'])
   2 [6, 4, 2, 3, 6, 4, 2, 4]

The following example shows the combined usage of a lambda expression

   1 map(lambda s:len(s)+1,['Always','look','on','the','bright','side','of','life'])
   2 [7, 5, 3, 4, 7, 5, 3, 5]